ARTICLE https://doi.org/10.1038/s42003-020-01169-9 OPEN Wireworm (Coleoptera: Elateridae) genomic analysis reveals putative cryptic species, population structure, and adaptation to pest control ✉ Kimberly R. Andrews 1 , Alida Gerritsen2, Arash Rashed 3, David W. Crowder4, Silvia I. Rondon 5, 1234567890():,; Willem G. van Herk6, Robert Vernon7, Kevin W. Wanner8, Cathy M. Wilson9, Daniel D. New1, Matthew W. Fagnan1, Paul A. Hohenlohe1 & Samuel S. Hunter 1 The larvae of click beetles (Coleoptera: Elateridae), known as “wireworms,” are agricultural pests that pose a substantial economic threat worldwide. We produced one of the first wireworm genome assemblies (Limonius californicus), and investigated population structure and phylogenetic relationships of three species (L. californicus, L. infuscatus, L. canus) across the northwest US and southwest Canada using genome-wide markers (RADseq) and genome skimming. We found two species (L. californicus and L. infuscatus) are comprised of multi- ple genetically distinct groups that diverged in the Pleistocene but have no known distin- guishing morphological characters, and therefore could be considered cryptic species complexes. We also found within-species population structure across relatively short geo- graphic distances. Genome scans for selection provided preliminary evidence for signatures of adaptation associated with different pesticide treatments in an agricultural field trial for L. canus. We demonstrate that genomic tools can be a strong asset in developing effective wireworm control strategies. 1 Institute for Bioinformatics and Evolutionary Studies (IBEST), University of Idaho, Moscow, ID 83844, USA. 2 Computational Sciences Center, National Renewable Energy Laboratory, Golden, CO 80401, USA. 3 Department of Entomology, Plant Pathology and Nematology, University of Idaho, Moscow, ID 83844, USA. 4 Department of Entomology, Washington State University, Pullman, WA 99164, USA. 5 Oregon State University, Hermiston Agricultural Research and Extension Center, Hermiston, OR 97838, USA. 6 Agassiz Research and Development Centre, Agriculture and Agri-Food Canada, Agassiz, British Columbia, Canada V0M 1A0. 7 Sentinel IPM Services, Chilliwack, British Columbia, Canada V2R 3B5. 8 Department of Plant Sciences and Plant ✉ Pathology, Montana State University, Bozeman, MT 59717, USA. 9 Idaho Wheat Commission, Boise, ID 83702, USA. email: [email protected] COMMUNICATIONS BIOLOGY | (2020) 3:489 | https://doi.org/10.1038/s42003-020-01169-9 | www.nature.com/commsbio 1 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-01169-9 he larvae of click beetles (Coleoptera: Elateridae), known as genomic level to different habitats, information that could help “wireworms”, pose a growing threat to agricultural crops in predict how populations will respond to control measures. In T 1,2 temperate and subtropical regions around the world . addition, genome scan approaches can be used to evaluate Wireworms feed on the sprouting seedlings, root, and stem tis- whether populations are evolving over time in response to sues of a wide range of economically important crops such as environmental change, such as changes in pesticide treatment wheat (Triticum spp.), corn (Zea mays L.), and potatoes (Solanum or IPM strategies; this evolution could lead to a reduction in the tuberosum L.). They typically live in the soil for 2–11 years, efficacy of treatment over time. However, despite the strong depending on the species and environmental conditions. Wire- potential utility of genetic approaches for the development of worms are becoming increasingly prevalent in croplands because effective IPM for wireworms, to our knowledge no population pesticides that effectively controlled them are no longer com- genetic or genome scan approaches have been conducted for mercially available due to concerns regarding human and envir- wireworm species. onmental health3,4. Current methods for controlling wireworms Here, we use genomic techniques to investigate putative cryptic are far less effective; the new generation insecticides are neoni- species, within-species population genetic structure, and local cotinoids, which only temporarily debilitate or repel the wire- adaptation of wireworms across a range of geographic scales in worms, without causing substantial mortality5–10. Moreover, the northwest US and southwest Canada. We focus on three alternative strategies involving cultural practices such as crop wireworm species that are resurging as pests: Limonius cali- rotation are not highly effective and are not feasible in many fornicus Mannerheim, L. infuscatus Motschulsky, and L. canus cropping systems, primarily due to the wide host range of this LeConte (Fig. 1)22,23. These species are endemic to the northwest pest complex1,2. US and southwest Canada, with L. californicus also occurring The lack of effective methods to control wireworm infestations throughout California24. Previous DNA barcoding studies found has led to a growing interest in developing integrated pest evidence for putative cryptic species within both L. californicus management (IPM) strategies for these species2,11. IPM utilizes a and L. infuscatus in this region17,18. We build on these studies by combination of biological, cultural, mechanical, and chemical expanding the geographic range of sampling, and by using high- controls, guided by knowledge of the biology and ecology of the resolution genome-wide sequence data from both nuclear and pest species12,13. This approach relies on the idea that different mtDNA markers. We sequenced and assembled the whole gen- pest species, or different populations of pest species, may respond ome for one L. californicus specimen, and generated genomic data differently to control measures; therefore, one of the first steps for from multiple individuals from each of the three species using effective IPM is to determine which species and populations are restriction site-associated DNA sequencing (RADseq), a techni- present in a given area. However, knowledge of wireworm species que that surveys thousands of loci across the nuclear genome25. composition and population structure is often limited, thereby We also generated mtDNA sequence data from multiple indivi- reducing the ability to design and implement IPM strategies. One duals using genome skimming, a technique that uses low- reason for this lack of information is that wireworm species in the coverage shotgun sequencing to retrieve mtDNA sequence larval stage are difficult to identify morphologically due to a lack data26,27. We identified genetic lineages within both L. californi- of clear distinguishing characters between species; only a limited cus and L. infuscatus that were highly divergent for both RADseq number of experts in the field are able to distinguish some spe- and mtDNA markers, potentially indicating the presence of cies. This limitation is exacerbated by the scarcely characterized cryptic species complexes. We also found evidence for fine-scale population structure of wireworms. within-species population structure for the L. californicus and L. Molecular approaches have strong potential to fill knowledge infuscatus species complexes, with genetically divergent popula- gaps and address practical challenges regarding wireworm species tions separated by as little as 25 km. We did not find genetically identification and population structure. DNA barcoding is one distinct groups within a single agricultural field for L. canus, but example of a molecular tool that relies on sequence data from genome scans for selection indicated that pesticide treatments mitochondrial DNA (mtDNA) markers to distinguish species14. may be driving adaptive evolution within the agricultural field. This technique can be particularly useful when morphology- The results of our study provide insight into the species com- based species identification is difficult. DNA barcoding efforts on position, population genetic structure, and genomic adaptation of wireworms from North America, Europe, and other regions have wireworms in the northwest US and southwest Canada that can demonstrated that portions of the cytochrome oxidase I (COI) assist in the development of effective IPM programs. gene and 16S ribosomal DNA (rDNA) are effective mtDNA 15–19 markers for distinguishing known species . These studies also Results found genetic evidence for the presence of wireworm species that Shotgun sequencing and genome assembly. Shotgun sequencing were previously undescribed due to the absence of distinct using genomic DNA from a single L. californicus specimen pro- morphological traits between them (cryptic species), indicating duced 160 million PE250 reads (80 Gbp) from Illumina HiSeq that the number of wireworm species may currently be 2500, 33 million PE300 reads (19.9 Gbp) from Illumina MiSeq, underestimated. 738,071 synthetic long reads with median length 1253 bp (1.47 Molecular approaches also offer effective methods in inves- Gbp) from Illumina SLR, and 1,750,383 reads with median length tigating within-species population structure and dispersal pat- 2400 bp from PacBio RSII (5.01 Gbp). Assembly produced terns. Genetic analyses investigating population structure 115,615 contigs with an assembled length of 1,072,695,639 bp, ideally use multiple nuclear markers. In the past, population and N50 of 19,399 bp. BUSCO analysis reported that 1476 of the genetic studies of non-model organisms typically used 10–20 1658 reference genes were complete (89%, 1387 single copy and nuclear microsatellite